metadata

library_name: transformers
license: mit
base_model: microsoft/speecht5_tts
tags:
  - generated_from_trainer
model-index:
  - name: speecht5_mehdi_new_as_try111
    results: []

speecht5_mehdi_new_as_try111

This model is a fine-tuned version of microsoft/speecht5_tts on the None dataset. It achieves the following results on the evaluation set:

Loss: 0.5467

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

learning_rate: 1e-06
train_batch_size: 4
eval_batch_size: 2
seed: 42
gradient_accumulation_steps: 8
total_train_batch_size: 32
optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: linear
lr_scheduler_warmup_steps: 100
training_steps: 3000
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
1.0666	1.7778	100	0.8627
0.8711	3.5556	200	0.7175
0.8314	5.3333	300	0.6804
0.8011	7.1111	400	0.6603
0.7676	8.8889	500	0.6402
0.7335	10.6667	600	0.6158
0.7093	12.4444	700	0.5889
0.676	14.2222	800	0.5793
0.6617	16.0	900	0.5743
0.664	17.7778	1000	0.5711
0.6516	19.5556	1100	0.5664
0.6478	21.3333	1200	0.5609
0.6445	23.1111	1300	0.5590
0.642	24.8889	1400	0.5601
0.6341	26.6667	1500	0.5585
0.6415	28.4444	1600	0.5584
0.6373	30.2222	1700	0.5533
0.6257	32.0	1800	0.5508
0.6311	33.7778	1900	0.5516
0.6201	35.5556	2000	0.5487
0.6257	37.3333	2100	0.5496
0.6304	39.1111	2200	0.5494
0.6177	40.8889	2300	0.5473
0.6235	42.6667	2400	0.5463
0.6202	44.4444	2500	0.5475
0.6191	46.2222	2600	0.5464
0.6188	48.0	2700	0.5442
0.6034	49.7778	2800	0.5452
0.6132	51.5556	2900	0.5453
0.6205	53.3333	3000	0.5467

Framework versions

Transformers 4.46.3
Pytorch 2.5.1+cu121
Datasets 3.2.0
Tokenizers 0.20.3